Failure mode and bimodal restenosis of drug-coated balloon in femoropopliteal intervention

Background: Pattern of DCB restenosis and associated outcomes in facing complex femoropopliteal lesions remain uncertain. Methods: Data were retrospectively collected from dual centers in Taiwan on patients who underwent treatment with DCBs for femoropopliteal lesions between 2013 and 2016. The restenosis pattern was categorized by the index-treated length. Clinical outcomes and time to DCB restenosis were retrospectively analyzed. Cox proportional hazards model identified restenosis predictors. Results: We recruited a total of 164 patients (91 men; median age 73 years) into the final analysis. The mean lesion length was 204.0 ± 109.2 mm. Of them, 45% total occlusions, 28% severe calcification and 15% in-stent restenosis were treated. Fifty-five patients have DCB restenosis (28 focal and 27 diffuse-occlusive patterns) over a 55-month follow-up. The median restenosis time emerged as a bimodal pattern with a significant difference between the diffuse-occlusive and focal restenosis group (225 vs. 484 days, P = 0.01). The 1-year patency rate after reintervention for DCB restenosis also was different between both restenosis group (29% vs. 65%, P = 0.017). The anticipated timing of escape for diffuse-occlusive or focal restenosis was 687 and 1068 days, respectively. Independent factors were lesion length (P = 0.049) for diffuse-occlusive restenosis and lumen gain of the popliteal artery for focal restenosis (P = 0.034). Conclusions: This study demonstrated time to DCB failure emerged as a bimodal pattern of distribution and associations of restenosis pattern to subsequent outcomes after the repeated intervention. Exemption from late catchup restenosis required 3-year observation instead of the 1-year mark for conventional treatment.